Method, apparatus and computer-readable medium for providing urinary stone information
AssigneeSTARLABS Co.
InventorsJun ho Lee, Min Soo Choo, Jin Kim, Chan Woo Kwak
Life of this patent
Estimating the term…
Abstract
The present invention relates to a method for providing urinary stone information, and more particularly, to a method for providing urinary stone information, capable of providing information necessary for urinary stone surgery by detecting a region where a stone is present from a plurality of tomography images by using a machine learning model, and automatically extracting information including a location and a size of the stone.
Claims10
A method for providing urinary stone information, which is performed in a computing system having at least one processor and at least one memory, the method comprising:
- extracting at least one stone region information by using a first machine learning model trained for each of a plurality of tomography images having mutually different tomography heights;
- extracting image-specific stone detail information related to a region actually corresponding to a stone in each stone region based on color information of a pixel in the stone region of the tomography image corresponding to the at least one stone region information;
- clustering the image-specific stone detail information for a same stone among the image-specific stone detail information of each of the tomography images;
- extracting stone-specific detail information including a size of each stone based on information including the stone region information and the image-specific stone detail information;
- and
- generating or providing the urinary stone information based on at least one of the tomography image, the stone region information, the image-specific stone detail information, and the stone-specific detail information according to a request of a user.
An apparatus for providing urinary stone information, the apparatus comprising:
- a stone region information extraction unit for extracting at least one stone region information by using a first machine learning model trained for each of a plurality of tomography images having mutually different tomography heights;
- an image-specific stone detail information extraction unit for extracting image-specific stone detail information related to a region actually corresponding to a stone in each stone region based on color information of a pixel in the stone region of the tomography image corresponding to the at least one stone region information;
- an image-specific stone detail information clustering unit for clustering the image-specific stone detail information for a same stone among the image-specific stone detail information of each of the tomography images;
- a stone-specific detail information extraction unit for extracting stone-specific detail information including a size of each stone based on information including the stone region information and the image-specific stone detail information;
- and
- a urinary stone information generation unit for generating or providing the urinary stone information based on at least one of the tomography image, the stone region information, the image-specific stone detail information, and the stone-specific detail information according to a request of a user.
Dependencies are inferred from the claim text.